Comparison of surface sensor and bone-fixed measurement of humeral motion

A common method of tracking humeral motion involves securing a thermoplastic cuff to the humerus with an electromagnetic sensor attached. The data on the accuracy of this technique are limited. This study addressed two questions: (a) How similar are surface and bone-fixed measurements of 3-D humeral rotations? (b) How similar are surface and bone-fixed measurements of 3-D humeral translations? Electromagnetic motion sensors were secured to a bone-fixed external humeral fixator, a surface humeral cuff, and the skin over the sternum and scapular acromion process. The 3-D data were collected during successive slow velocity (10-20degrees/second) repetitions of humeral active-assisted scapular plane abduction, sagittal plane flexion, and internal/external rotation with the arm adducted. Root mean square errors of surface measures compared to bone-fixed angular and translational values were calculated, and paired t-tests were computed between the two methods. Root mean square errors for humeral rotations ranged from 1degrees(1%) for humeral elevation during scapular plane abduction to 7.5degrees(9%) for humeral internal/external rotation. Peak errors were under-representations of 5.7degrees for internal/external rotation during scapular plane abduction and 15.6degrees for internal rotation with the arm adducted at the side. Average translation errors ranged from 0.1 to 2.1 mm. Data from this study suggest that dynamic measurement of humeral motion with a surface humeral cuff sensor can be performed for certain slow velocity motions with root mean square errors less than 8degrees. Caution is called for when interpreting internal/external rotation values, which were underrepresented. Results may vary with one's age, weight, or general physical condition, with different velocities of movement, or with different movements.